The combustion synthesis of fullerenes was described in Nature, 352, 139-141, 1991 and U.S. Pat. No. 5,273,729, both of which are incorporated herein by reference. Data were presented for fullerene yields from subatmospheric pressure premixed laminar benzene-oxygen flames operated under different conditions of pressure, feed velocity (and hence temperature, which depends on feed velocity), diluent identity and partial pressure, carbon/oxygen ratio, and distance or residence time from the burner. Different versions of the same basic method can be envisioned depending on burner design and other process features for which many variations are possible. In general the method involves the operation of a sooting flame at pressures less than atmospheric and other conditions such as carbon/oxygen ratio, feed velocity, and concentration of inert diluent gas chosen so as to be suitable for fullerenes production. Condensable material containing fullerenes is collected from within the flame or from the effluent gas. Subsequent to the original work on combustion synthesis of fullerenes, which involves the use of premixed flames, the use of non-premixed or diffusion flames has also been found effective.
Methods for combustion synthesis of fullerenes known in the art use burner stabilized flames suitable for laboratory scale studies. The conventional combustion method for fullerene synthesis is typically carried out using low-pressure laminar premixed flame burners in which the flame is stabilized on a water-cooled burner plate. The fuel and oxidant are usually premixed upstream of the burner and fed through openings in the drilled or sintered metal plate. In some cases the fuel and oxidant are fed separately through alternating fuel/oxidant openings in the burner plate and mix together immediately downstream of the exit of the plate, either prior to entering the flame and hence giving a premixed flame or partially before and partly after entering the flame and hence giving a partially premixed and partially non-premixed or diffusion flame. In all these cases the flames are stabilized a short distance off the surface of the burner plate, and heat loss to the water-cooled burner lowers the temperature of the flame such that the flame speed, defined as the velocity of propagation of the flame into the unburned fuel-oxidant mixture, just matches the incoming velocity of the mixture being fed to the flame. The stabilization of the flame at the particular short distance from the burner where the heat loss into the burner is just the right amount to lower the temperature the correct amount to allow the flame speed to match the incoming velocity places considerable constraints on the feed velocity and other conditions required to achieve and maintain the stable flame and thereby severely limits the degrees of freedom available for configuring and optimizing the flame for particular desired performances in fullerenes synthesis. Also the rate at which the fuel/oxidant mixture can be fed to the burner, and hence the throughput of the system, is severely limited by the flame speed of the mixture. A different design that decouples flame stabilization from product formation, thereby giving more flexibility for operating and optimizing the process, while also allowing larger throughputs is highly desirable.
Low-pressure non-premixed or diffusion flames have also been used to form fullerenes. In these flames the fuel and oxidant are fed separately through different ducts or passage ways in the burner, and they mix together in a flame zone stabilized between the two streams. Both laminar and turbulent versions of these flames have been used to form fullerenes. Although fullerenes can be found within the flame zone, they tend to be destroyed by oxidation in flames where the fullerenes exit through the oxygen side of the flame zone. In flames where the fullerenes exit through the fuel side of the flame zone, they are formed in the presence of, and react with, large concentrations of soot and polycyclic aromatic hydrocarbons. Thus neither of the nonpremixed or diffusion flame types have proven interesting for practical fullerenes production.
There are many improvements possible to these types of flames for the efficient and economical production of commercial quantities of fullerenes. Notably, conventional fullerene forming flames have limited flow rates, the maximum flow rate being determined by the flame speed for the given fuel/oxidant system and temperature, with temperature being determined by heat losses from the flame. Control of reaction variables, particularly temperature, is limited, since setting flow rate and heat loss determines the temperature and also the radical flame intermediate concentrations and the time at which the fuel spends in the oxidation region. The oxidation region is important in determining the concentrations of key intermediates. The temperature and residence time necessary for oxidation reactions are not necessarily the conditions which are optimal for fullerene formation, and it would be desirable to have a degree of control over the conditions of reaction so as to favor fullerene formation, without being too constrained by the necessities of flame stabilization. A reactor that achieves higher throughputs would also be preferable to current methods, due to future higher volume requirements of fullerenic product.
By “fullerenic product”, as used herein, it is meant material consisting of or including one or more of the following three types of material, structures or particles: (1) fullerene molecules such as C60, C70, C84, etc., fullerene molecules containing another atom or atoms inside or outside the fullerene cage or one or more functional groups; (2) fullerenic nanostructure or closed cage structures made up of five-member and six-member, and in some cases seven-member, carbon rings having at least one dimension on the order of nanometers, such as but not limited to single or multilayered nanotubes and nanoparticles as defined in U.S. Pat. No. 5,985,232, columns 3 and 4, and (3) fullerenic soot consisting of spheroids or spherules of carbon made up of curved carbon sheets or layers which have substantial fullerenic character. The spherules have dimensions similar to conventional carbon black and thermal black, that is, in the range of 5 nm to 1000 nm. Fullerenic character is noted by the presence among six-member and sometimes seven-member rings of five-member carbon rings which result in curved sheets of carbon.
Another important requirement for the commercial manufacture of fullerenes is an increase in the yield of fullerenic product for a given quantity of carbon fed as fuel. Currently, the highest reported yields are ˜0.5% of total carbon fed. A combustion reactor allowing a greater degree of control over the combustion reactions and better maintenance of conditions that favor or promote larger fullerene yields would be preferable to conventional systems.
Jet-stirred reactors which approximate a well-mixed combustion reactor have been used extensively in experimental combustion work, beginning with Longwell (Ind. Eng. Chem., 47, 1634, 1955)), but fullerenes have never been synthesized in a well-mixed combustion reactor. Neinninger (Proceedings of the Combustion Institute 20:473-479, 1984) and Dagaut, et al (J. Phys. E.: Sci. Instrum. 19, 207-209, 1986) are recent designs. Current jet-stirred reactors used in combustion are designed for atmospheric pressure conditions with turbulent gas flows from the jets and turbulent flow conditions within the reactor. Turbulence greatly increases the rate of diffusion and thus enhances micro-mixing. Turbulence enhanced micro-mixing is the primary method relied upon to create a well-stirred condition in existing jet-stirred reactors. Practically significant amounts of combustion generated fullerenes have only been found in low-pressure conditions (10-100 torr), for which the previous jet-stirred reactor designs would be inadequate in the amount of mixing and recycle accomplished because of the much lower densities and hence lower Reynold's numbers at the low pressures of fullerene formation, and turbulence is not adequate as the primary method of back-mixing or recycle. A jet-stirred reactor that can accomplish adequate recycle or back-mixing at low pressures, without relying primarily on turbulence enhanced micro-mixing is needed.
Lam (Proceedings of the Combustion Institute 22:323-332, 1988) describes a jet-stirred reactor/plug-flow reactor system. This system was used as a laboratory tool for atmospheric pressure studies of polycyclic aromatic hydrocarbon and soot formation using primarily ethylene as the fuel. Total residence times in this system were of the order 5 ms in the jet-stirred reactor and 15 ms in the plug-flow section. Due to the high flow rates and low residence times, flow in the plug-flow section was turbulent, and due to the low residence times, external heating was not required to assume that temperature was isothermal in the plug-flow section. Such a coupled well-mixed/plug-flow reactor system has not been used at low pressures nor at residence times greater than 20 ms, and fullerenes have never heretofore been synthesized using a well-stirred/plug-flow reactor configuration.